Antimicrobial peptides

We focused our immunohistochemical studies on several parameters of the innate immune system, in order to understand whether or not parts of the innate system participate in the mucosal defence of the mammalian esophagus epithelium.

Although constitutive, this system is activated by the presence of microorganisms and their products, providing a rational for a potential bio defence strategy (NEISH 2009). Besides our aim to detect receptors of the innate immune system expressed by epithelial cells (TLRs) or on sentinental cells (LCs), our intent was to detect substances produced and secreted by the epithelial cells (APs, ß-glucan receptors, and lysozyme).

Mucosal defence strategies in the human esophagus

Information about esophageal mucosal defence is sparse, even in humans.

Nevertheless, HOPWOOD (1995) demonstrated several interesting aspects of esophageal defence, e.g. various defence mechanisms can be elicited by reflux damage. The author argued that chemical damage wrought by proteinases is countered by cell turnover, intercellular barriers and pH management.

Furthermore, the presence of epithelial growth factor (EGF) is an added defence feature that acts when ulcers are formed. For our study, the evidence of an

“intercellular barrier” is of particular interest. HOPWOOD (1995) demonstrated that the intercellular space between the cells in the stratum spinosum was not empty, as it appeared to be in electron microscopy, but showed the presence of

“mucosubstances”. However, during his studies the characteristics and functions of this material remained unclear. In contrast to these findings, we confirmed the presence of APs as intercellular material between the layers of the esophagus epithelium of the nine mammalian species studied. This might be due to the fact that the substances accumulate in large amounts between cells of the stratum spinosum only in the human esophagus (HOPWOOD 1995), but are generally present in smaller amounts in other layers. In our study we used highly sensitive

IHC to detect APs, so that it was possible to give a detailed description of their distribution in the esophagus epithelium.

Additionally, we provide the first artefact-free electron microscopical pictures (cryo-SEM) of the surface of the esophagus epithelium, thus verifying the existence of a one-layered microbial colonisation in all species studied. To be able to tolerate and benefit from these microorganisms, the eukaryotic host must monitor the microbiota and control their number and composition. One possible way to control colonisation with commensal and potentially pathogenic bacteria is the production of APs, which are part of the first line of defence (INAMOTO et al. 2008; NEISH 2009).

As APs are found in various epithelial tissues (DALE and FREDERICKS 2005;

BRAFF and GALLO 2006) and due to their antimicrobial and immunoregulatory activity, scientists perceive APs as promising candidates for new therapeutic drugs, also in veterinary medicine (LINDE et al. 2008; SANG and BLECHA 2009).

In accordance with the sites expressing APs, our discussion focuses on two other organ systems. To start out, an analysis of the AP expression pattern in other parts of the digestive tract is of interest. As a matter of fact APs are involved in the control of the physiological microbial flora and mucosal homeostasis (BEVINS et al. 1999; IIMURA et al. 2005; LOTZ et al. 2007; SALZMAN et al. 2007;

MUKHERJEE et al. 2008; VELDHUIZEN et al. 2008). For example, the production of APs (mainly α-defensin) in the small intestine of humans is restricted to Paneth cells located in the intestinal crypts, whereas in the large intestine CAT and ß-defensins contribute to the antimicrobial barrier (LOTZ et al. 2007). Interestingly enough, reduced numbers of Paneth cells and a decrease in the α-defensin production were noted in Crohn`s disease patients (WEHKAMP et al. 2005).

Thereafter the specific construction of the esophagus epithelium has to be taken into account. In most herbivorous and omnivorous species the esophagus shows a more or less keratinised stratified squamous epithelium, which is not really comparable to the one of the epidermis of the skin. The latter organ presents a reliable mechanical, as well as an immunological barrier to infection, displaying a considerable capacity of innate immunity. Due to constant attacks of microorganisms and remarkable mechanical strain, research focussed on the expression of APs in the epidermis (OREN et al. 2003; BRAFF et al. 2005; ELIAS

2005; ELIAS and CHOI 2005; BRAFF and GALLO 2006). Whenever the occurrence of this substance group is discussed, it is of great interest whether the APs are produced constitutively or inducible at the site of inflammation. BRAFF et al. (2005) found that CAT is constitutively expressed by neutrophils in the skin and inducible in keratinocytes in response to infection. In contrast to this, ELIAS and CHOI (2005) argued that hBD-2 and -3, and CAT are expressed at low levels in unperturbed skin, but occur in higher levels in healing wounds and in inflammatory dermatoses. Our results demonstrated that hBD-2 and -3 and CAT are expressed constitutively in the esophagus epithelium of the species studied. It is inconceivable that all animals were infected, and thus the constitutive production of the considered APs is necessary to maintain the homeostasis of the esophagus epithelium. None of the animal species revealed negative results, only the reaction intensity differed between the species. Due to our findings we agree with ELIAS and CHOI (2005) and can transfer their findings demonstrated for the skin to a certain extent to the esophagus epithelium of the domesticated mammals studied.

Nevertheless, additional experiments analysing the influence of infection by certain pathogens need to be conducted, whereby the up-regulation of APs could be a likely effect.

Comparison of expression sites in the esophagus epithelium

In the skin, APs are normally stored in so-called lamellar bodies (LBs) or membrane coating granules (MCGs), respectively, but were also found in the intercellular space (OREN et al. 2003; ELIAS and CHOI 2005). While the glycolipid contents of the LBs are well known, the additional AP contents were a surprise.

OREN et al. (2003) gave a detailed description on the expression sites of hBD-2 in the epidermis. They performed immunogold labelling and showed that hBD-2 is predominantly located in the spinous and granular layers. Dense labelling was demonstrated for the LBs of the keratinocytes and some additional staining was seen in the intercellular space. In contrast, ELIAS and CHOI (2005) described a different expression pattern of hBD-2 and additional hBD-3 and CAT in the epidermis. Conducting IHC, they showed that all three APs were predominantly localised in the outer epidermis. At this site they appeared in LBs and in the intercellular spaces of the stratum corneum. The authors argued that the APs are

perfectly positioned to intercept pathogenic microorganisms when they attempt to penetrate the epidermis in between the corneocytes. With regard to our own findings from the esophagus epithelium, we could not corroborate the results of OREN et al. (2003) and ELIAS and CHOI (2005). One contrasting finding was the observation in all species and APs studied, that constantly strong to very strong reaction intensities occurred in the cells of the stratum basale. Due to the fact that this layer has the highest rate of proliferation, we presume a direct connection to AP production intensities. With regard to the three nutrition types studied, a statistically significant difference could be shown for hBD-3 in the stratum basale.

Interestingly enough, the carnivorous cat exhibited a slightly stronger reaction in comparison to the herbivorous and omnivorous species analysed. The increased production of hBD-3 in the feline stratum basale could be considered a reaction due to diminished preventative mechanical protection properties of the esophagus epithelium. The feline esophageal epithelium is only weakly keratinised, and therefore cannot function as a reliable mechanical barrier against invading microorganisms. Thus, the epithelium has to be protected by larger amounts of APs. Supporting evidence for this theory is provided by the results observed for the stratum corneum. What is more, statistically relevant differences were found in this layer when comparing herbivorous and omnivorous species with the carnivorous cat. This time hBD-2 was expressed in greater amounts in the carnivorous cat. Moreover, hBD-3 was expressed at a higher level in the cat, even though the difference between the nutrition groups was not obvious from the statistical point of view. Consequently, the feline esophagus epithelium is intensively protected by ß-defensins. Due to its reduced mechanical protection by the weakly keratinised epithelium and the constant influence of microorganisms, ß-defensin production is remarkable and has to fulfil a distinct protective task.

Presumably hBD-2 is more effective on the epithelial surface, and hBD-3 functions in the more basal part of the feline esophagus epithelium. This explanation and the exact amount of APs produced remains to be examined in future experiments (for example by the application of densitometry). On the one hand, these observations are to some extent astonishing, as one would expect the herbivorous species to exhibit the strongest reactions, due to their rough-textured food and the resulting high mechanical strain. On the other hand, our findings showed the influence of other epithelial properties, such as a regulated AP production. A more keratinised

epithelium protects the mucosa from invading microorganisms mechanically, and a less keratinised epithelium is protected by an increased production of ß-defensins.

It remains to be elucidated why the reaction intensity of all three APs decreased from the stratum basale towards the lower and upper stratum spinosum. One explanation could be that APs are stored in vesicles, comparable to the LBs (MCGs) of the epidermis (OREN et al. 2003; ELIAS and CHOI 2005), and are not detectable by the primary antibody, due to the compact packing. The idea of substance transport in vesicles corresponds with findings of INAMOTO et al.

(2008), who investigated the epithelial response towards the attachment of indigenous bacteria in the small intestine of the rat. One of their observations was that defensins and lysozyme are transferred via cellular vesicles into the lumen, where they fused with the membrane of an invaginated bacterium. However, in some cases an increase in reaction intensity could be observed towards the stratum granulosum of the esophagus. Concerning hBD-2, this was true for the goat, sheep, pig, mouse and rat. Regarding hBD-3, the phenomenon was demonstrated for cattle, sheep, pigs and mice, and for CAT an increase was observed in the ovine and porcine esophagus epithelium. Analysing the species that showed the increase mentioned makes the existence of a correlation between these results and the nutrition type a realistic feature. All of the species are herbivorous or omnivorous and because of the enormous mechanical strain on their esophagus epithelium, in comparison to the esophagus of the carnivorous cat, APs have to be stored in the stratum granulosum. This “layer” of APs in the latter stratum probably indicates some kind of barrier function. An intriguing finding is that such a “barrier” was constantly detectable in the porcine esophagus epithelium. As shown by the analysis of the structure of the esophagus epithelium, the pig revealed a rather loose and varying organisation of the stratum corneum.

Such feature may result in an increased penetration of microorganisms. This specific structural aspect explains the necessity of a protective layer of APs. Due to the fact that the esophagus epithelium of the cat is protected by a layer of hBD-2 covering the surface, the storage of APs in the stratum granulosum is not necessary in this species. For CAT, the strongest reaction intensity was seen in the pig, mainly in the stratum basale. One explanation could be that the commensal flora of the porcine esophagus especially triggers this AP production.

LINDE et al. (2008) showed in in-vitro experiments that CATs are produced by the

porcine organism and demonstrated its high effectiveness against: E. coli, Salmonella typhimurium, Staphylococcus aureus, Actinobacillus pleuropneumonia, Pseudomonas aeruginosa, and Candida albicans. It remains to be elucidated via microbial differentiation, whether these microorganisms also belong to the commensal flora of the porcine esophagus surface.

LINDE et al. (2008) showed in-vitro activity of equine ß-defensins against the following microorganisms: Corynebacterium spp. and Staphylococcus intermedius.

These two Gram-positive cocci were not found in the microbial flora of the esophagus of the horse (MEYER et al. 2009). Furthermore, LINDE et al. (2008) described in-vitro activity of CAT against the following bacteria: Escherichia coli, Streptococcus equinus, Klebsiella pneumonia and Serratia marcescens. In contrast to the afore-mentioned results for ß-defensins, this spectrum of antimicrobial activity is more consistent with the findings made by MEYER et al.

(2009) concerning the actual microbial status of the equine esophagus. These authors demonstrated a high occupation with the Gram-negative facultative anaerobic rods E.coli. However, E.coli is an ubiquitous bacterium and cannot be made out as specific for the equine species or generally explain the expression of CAT in the equine esophagus. The results indicate that the characteristic microbial flora of the esophagus of the different domesticated mammals still needs to be analysed more intensively.

With regard to our own findings and the literature studied, it can be concluded that the diverse repertoire of APs is most likely a key factor in allowing the esophagus mucosal surface to maintain homeostasis concerning the diverse colonising bacterial populations.